Connect pp-import to the preprocessor grammar #3575
Conversation
Clarify how the grammar pp-import relates to control-line.
| @@ -671,6 +676,9 @@ | |||
| \indextext{header unit!preprocessing}% | |||
| \indextext{macro!import|(}% | |||
|
|
|||
| \pnum | |||
| An \grammarterm{import-line} shall be of the form | |||
There was a problem hiding this comment.
For "of the form" constraints, we usually don't show another non-terminal, but just the grammar sequence that needs to be satisfied. That said, I think the overall change is a net improvement.
There was a problem hiding this comment.
I noticed that, but the subsequent text refers to things like "all three forms of pp-import" etc. So I thought it nice to keep that name. I suppose we could change it to "all three forms above" and "the second form", etc. Not sure if that would be better though.
There was a problem hiding this comment.
We do the same thing in [class.conv.fct], which is much more egregious than this (it lists three productions and only means the first one). Though I do wonder if we can do better here; perhaps we could list pp-import as a control-line production (perhaps renaming it to import-line), and change [cpp.pre]/2 to say that a text-line is only matched if the line is not a preprocessor directive. (That way, an import directive that's not a pp-import would simply not match the preprocessor grammar, just like an unmatched #if or #endif.)
jensmaurer
added
the
changes requested
jensmaurer
added
the
needs rebase
|
Please rebase and force-push. |
|
@daveedvdv: Could you please take @zygoloid's suggestion into consideration? And rebase? |
|
We're not sure if this is editorial. We agree that there's a bug in the wording, but the detailed resolution needs oversight, too, so could you please raise this proposal as a CWG issue? |
|
Fixed by 862fe4b (Feb 2020) |
Clarify how the grammar pp-import relates to control-line.